Received-SPF: pass (google.com: domain of linux-wireless+bounces-8072-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Message-ID: <f939f9dfb467d7a1daaf1e69b56f8599ef3eceeb.camel@sipsolutions.net>
Subject: Re: [PATCH] wifi: mac80211: add input validation to
 sta_stats_decode_rate()
From: Johannes Berg <johannes@sipsolutions.net>
To: James Dutton <james.dutton@gmail.com>, linux-wireless@vger.kernel.org
Date: Mon, 27 May 2024 07:41:22 +0200
In-Reply-To: <CAAMvbhGyheFdWSrDzM_i10n9s06n3G2wX6O_S68VUZyP-a9p+A@mail.gmail.com>
References: 
	<CAAMvbhGyheFdWSrDzM_i10n9s06n3G2wX6O_S68VUZyP-a9p+A@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.50.4 (3.50.4-1.fc39) 
Precedence: bulk
MIME-Version: 1.0

On Mon, 2024-05-27 at 00:43 +0100, James Dutton wrote:
> Validation is required as a result of parameters derived from
> received wifi packets.

I don't think I fully agree with that. First of all, this data is never
actually directly derived from the wifi packet (certainly not any
pointers or the band enum!), even the PLCP contains different encodings.
Thus there's always already a translation in driver or firmware.

Now of course we shouldn't trust firmware either, but even then there
are a lot of places, I'd think this is better done at the driver level.

johannes